Amazon reviews as corpus for sentiment analysis using machine learning

Abstract

Most of the users today, provide their reviews on the various products on the Amazon website. The reviews provided by users are usually compact and demonstrative. For this reason, it becomes an affluent source for sentiment analysis. The objective of the paper is obtaining comparisons of the working of four standard Machine Learning algorithms for classifying the sentiments of the considered Amazon product reviews dataset. In our work, we determine the performance of these algorithms, that is, how accurately can they classify the sentiment of an unknown review. The paper provides a brief insight on sentiment analysis and the comparison of the performance of the considered algorithms of the classification of the sentiments based on several performance metrics.